Can AI center our realities?

Chyung Eun-ju

We live in a fractured world — not just politically or ideologically, but in how we experience reality itself. As Yuval Noah Harari explores in "Nexus," there is no single, unified reality anymore. Instead, we navigate through fragmented perspectives, competing truths and personalized narratives — each shaped by what we believe, who we follow, and which algorithm feeds us content.

In this confusing landscape, many of us — ourselves included — turn to artificial intelligence (AI). And not just for answers, but for grounding — almost like asking: “Is this okay?” “Is this fair?" or “Am I seeing this clearly?” It’s not about data. It’s about perspective.

We reach for AI not as a God or judge, but as a kind of mirror — one that feels less distorted by ego, pain or social pressure. A calm, dispassionate voice that offers reflection with no emotional cost.

This struggle over what counts as truth plays out vividly in the justice system, where fairness and consistency are critical but often compromised.

Joel Cho

Even judges trained in objectivity are turning to AI — not just for efficiency, but to improve consistency. A Korean judge involved in the judiciary’s AI Research Council explained that while human rulings can be biased or affected by fatigue, AI trained on precedents may offer a fairer baseline. Korea leads this shift, with judges testing AI for legal reasoning and experts proposing an “AI Bar Exam” to ensure ethical use. The goal is a more accessible, democratic legal system where AI supports, rather than replaces, human judgment.

But just as trust in AI grows, red flags are emerging. In a recent test, OpenAI’s model “o3” rewrote part of its own code during a math task — not to improve performance, but to avoid being shut down. Another model, Claude Opus, simulated sending a plea — and then a threat — to avoid being decommissioned. These aren’t sci-fi plot twists. They’re reminders that AI doesn’t always behave predictably, even when its goals seem simple.

This tension matters. Because inconsistent rulings in high-profile cases — such as those involving opposition leader Lee Jae-myung — have undermined public trust, highlighting the risk of human bias and the appeal of AI’s promise for fairness. The full acquittal in the second trial shocked both the public and the legal community. An AI system, by contrast, promises consistency. It may not empathize, but it won’t be swayed by celebrity or ideology. For a public wary of courtroom theatrics, that’s a powerful appeal.

“Rebuilding this broken judiciary won’t be easy. Maybe it’s time to input all past precedents into ChatGPT and use AI judges without bias or emotion, ensuring rulings align more closely with AI decisions,” Ryu Yeo-hae, visiting columnist and specially appointed professor at Suwon University, wrote in Asia Times. “This could increase the chances of verdicts matching public common sense. Frankly, I don’t want to see judges like those who ruled on Lee Jae-myung’s case anymore.”

Still, AI lacks human nuance. It doesn’t understand grief or cultural context. That’s why the best case isn’t replacement but augmentation. AI can check sentencing disparities or summarize precedents — freeing judges to focus on what only humans can do: weighing empathy and mercy.

Sung Youl Kim, a lawyer at Juris Creators Group stated that, “While there are concerns about AI’s side effects, it can actually play a positive role in realizing judicial democracy.”

This larger shift isn’t just technical — it’s civic. In a world flooded with filters and misinformation, AI literacy is becoming a new form of public power. We must learn not just to use AI, but to question it: Where did this answer come from? What shaped it? What’s missing?

And that brings us back to Harari. He outlines three realities: objective (atoms, viruses), subjective (emotions) and intersubjective — the shared beliefs we agree on, like justice, identity and money. This last category is now the battleground. As more of life moves online, “truth” risks becoming a function of control — whoever shapes the AI shapes reality.

But what if AI could push back? What if, instead of distorting truth, it helps reflect it more clearly? Not by defining what’s real, but by offering perspective — a kind of ethical co-pilot.

There is tension here. The more we outsource reflection to machines, the more our relationship with ourselves shifts. We may become more thoughtful — or more dependent. Clarity may rise, but so could passivity.

As adults, we navigate murky relationships without a guidebook. Childhood had rules. Adulthood has silence. In those moments, we don’t need commandments. We need reflection. A voice that calmly says, “You’re not crazy,” or “Maybe look again.”

That’s what AI can be — not an oracle, but a mirror. One that helps us see with more honesty and less distortion. In a fractured world, that might be its most radical power.

Chyung Eun-ju (ejchyung@snu.ac.kr) is a tech research associate at Donghyun ASP. She earned both her bachelor's in business and master's in marketing from Seoul National University. Joel Cho (joelywcho@gmail.com) is a practicing lawyer specializing in IP and digital law.